Re: Data Replication

Melvin Mariney wrote: > If I understand your question, I believe you are asking if you can > make the secondary server the primary when the original primary has > crashed and the answer is yes. There are three parameters in 7.1 HDR > that controls what happens when a the primary server fails. DRAUTO=0 > means that the secondary server stays in read-only mode and you have > to manually switch it to be updateable, DRAUTO=1 means that the > secondary is automatically switched to updateable or standard and > switched back to secondary once the primary is back up and replication > restarted, and DRAUTO=2 means that the secondary is automatically > switched to be updateable and is also made the primary server after > replication is restarted and the original primary is switched to be > the secondary server. We have four DBs that use replication and have found that the auto-recovery feature (DRAUTO = 1 or 2) does not work in our environment. DRAUTO = 0 is a valued feature for these mission critical DBs however. The configurations are HP K400 and I40, HP-UX 9.04 and 10.0, with OnLine 7.11.UC1X2. If you are using replication I highly recommend this release of the engine as it has several related bug fixes in it. Some fragmentation bugs were also addressed. These fixes were to have been rolled into 7.12 but I have not installed it yet. To answer the earlier question one of our replicated pairs is connected over a WAN and recovery to the secondary is fairly quick. Again this is done manually and the apps (Client PC's) must be "re-homed", essentially a cycling of the app. The need to re-orient the applications is due to our chosen architecture and in other designs can (and should) be automated. Enjoy, Bill